[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference ulysse::rdb_vms_competition

Title:DEC Rdb against the World
Moderator:HERON::GODFRIND
Created:Fri Jun 12 1987
Last Modified:Thu Feb 23 1995
Last Successful Update:Fri Jun 06 1997
Number of topics:1348
Total number of notes:5438

112.0. "4GL's and VAXset?" by KYOA::HANSON (I am NOT Drain-Bamaged.) Fri Apr 01 1988 17:56

Greetings!

I've also submitted this note to other conferences; my apologies for any
redundant redundanies.

Recently, I helped to conduct a rather large VAXset seminar at the local
office, and a few interesting questions came up, particularly with respect
to VAXset products being used in a 4GL development environment.  The
bottom-most question is :

 "Which VAXset products, if any, can be effectively used when developing
  applications either in ADABASE/NATURAL or ORACLE ?"

---------------

*MY* questions for the RDB_Competition folk might be :

  "Do the products ADABASE/NATURAL and ORACLE have features such as a
   common data dictionary, a source librarian or catolouge function; 
   Are compiles/generations done from an upper level as might be seen
   using MMS?  Do they generate 3GL code (like Cobol) or are they run-
   time, like with RALLY ?   Etc, etc.

---------------

My impression is that if editing of "source" is done via ASCII text, then
LSE can be used with a custom-tailored template.

But, SCA may not be applicable.

PCA might be used if these products are generating a 3GL type of platform,
and DTM would almost undoubtably give some benefit.

This is the general line of questioning.  Those that know both the VAXset
and the 4GL environment in question might be able to map the functionality
far better than I can.

Any thoughts?

Thanks,
Bob Hanson
T.RTitleUserPersonal
Name
DateLines
112.1DoubtfulQUILL::BOOTHA Career in MISunderstandingSat Apr 02 1988 00:4038
    Let's get real!   
    
    For Adabas/Natural, the user would have to write a template for
    an entirely new language, Natural. Every time new elements were
    introduced, the user would have to change or expand the template.
    While this could be done, I doubt that many customers would be willing
    to expend the effort.
                      
    Adabas/Natural uses PREDICT as its active Dictionary. 
    CMS could be used to store source, but I'm not sure that the Natural
    compiler could then find the code to compile since CMS stores original
    code first, then only the editing changes to the original. I'm not
    sure that Natural could decipher the format.
                      
    MMS might be useful, but the advantage of a 4GL is that it is
    non-procedural. Using a procedural module map compiling technique
    runs philosophically counter to 4GL advantages.
                      
    Since Adabas is doing the compiles, SCA is useless.
                      
    Without MMS and CMS for building and storing, would you really want
    to try DTM. I can't imagine PCA working with the Adabas utilities
    doing the compiles.
                      
    Oracle is much the same. It uses SQL*Dictionary as its active
    dictionary.       
                      
    SQL*Forms is a 4GL. It too would require a brand new template. Would
    the template and LSE give the user advantages over the screen-based
    system used in SQL*Forms? Not likely.
                      
    One of the problems with 3rd party databases is that they cost us
    huge amounts of downline software business like VAXset. But I think
    that retrofitting VAXset for each particular database is an exercise
    in futility.      
                      
                      
    ---- Michael Booth
112.2works great as long as you don't change itCOOKIE::JANORDBYMon Apr 04 1988 22:1922
    Being new to Digital and not yet master of the TLA's (Three Letter
    Acronyms), I my stumble a little here, but let me give it a shot.
    NATURAL is written and positioned as a 4th generation application
    development system, of which a major part is the 4th generation
    language, also called NATURAL. A customer attempting to use some
    of Digital's development products to enhance NATURAL would be
    abdicating the benefits that he bought NATURAL for in the first
    place. NATURAL contains its own editors (not language sensitive),
    its own screen generator, a library management system (not very
    good, but for both source and object), and is very integrated with
    the dictionary, Predict. Needless to say, source, object and dictionary
    information is stored in ADABAS (no E) and is verrrry efficient
    in terms of retrieval, compression, etc. But, if any one of these
    functions is managed by something other than ADABAS/NATURAL, the
    integration falls to pieces. Some parts of the system simply would
    not know where other pieces are. As things like SUPERNATURAL, NATURAL
    ELITE, NATURAL SECURITY, and CONNECT are added to the environment,
    the interdependencies magnify geometrically. This gets Digital in
    two ways: 1) Rdb and associated products are locked out and 2) since
    every one of the above mentioned products is written in NATURAL,
    each will run on IBM without changing a thing. Your customer now
    has an open door to try a 9370 or worse.
112.3I'll sign this oneCOOKIE::JANORDBYMon Apr 04 1988 22:239
    one more note. NATURAL contains a language, among many other
    facilities, just like COBOL does. It compiles to object code, but
    does require runtime modules (as does COBOL, at least in the IBM
    world). It does not generate a 3gl, that would be like COBOL generating
    assembler before compiling. Most NATURAL functions perform fairly
    well on the VMS platform, though the new version has many performance
    problems on the IBM platform.
    
    Jamey Nordby
112.4VAXset and RdbTHATIS::SIMPSONSteve Simpson, Reading EnglandTue Apr 05 1988 15:392
Whilst on the subject of VAXset, don't forget that you shouldn't sell the LSE 
compile/review feature with Rdb pre-compilers.